Towards a hybrid recommendation approach using a community detection and evaluation algorithm

Abstract

In social learning platforms, community detection algorithms are used to identify groups of learners with similar interests, behavior, and levels. While, recommendation algorithms personalize the learning experience based on learners' profile information, including interests and past behavior. Combining these algorithms can improve the recommendation quality by identifying learners with similar needs and interests for more accurate and relevant suggestions. Community detection enhances recommendations by identifying groups of learners with similar needs and interests. Leveraging their similarities, recommendation algorithms generate more accurate suggestions. In this article, we propose a novel approach that combines community detection and recommendation algorithms into a single framework to provide learners with personalized recommendations and opportunities for collaborative learning. Our proposed approach consists of three steps: first, applying the maximal clique-based algorithm to detect learning communities with common characteristics and interests; second, evaluating learners within their communities using static and dynamic evaluation; and third, generating personalized recommendations within each detected cluster using a recommendation system based on correlation and co-occurrence. To evaluate the effectiveness of our proposed approach, we conducted experiments on a real-world dataset. Our results show that our approach outperforms existing methods in terms of modularity, precision, and accuracy.